This completes the circle, complementing -lto-embed-bitcode
(specifically, post-merge-pre-opt). Using -thinlto-assume-merged skips
function importing. The index file is still needed for the other data it
contains.
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Minor comments.
clang/test/CodeGen/thinlto_embed_bitcode.ll | ||
---|---|---|
25 | Can we do this clang invocation and diff later on so you don't need to redo the opt and llvm-lto steps below? | |
llvm/include/llvm/LTO/LTOBackend.h | ||
59 | Note this is for distributed ThinLTO. | |
llvm/lib/LTO/LTOBackend.cpp | ||
75 | Nit on name/comment. It isn't just thinlto importing (also things like linkage type changes). I'm not sure of a concise name to represent this though. Because we do apply some ThinLTO optimizations from the index later (WPD). Maybe just expand the description to something like "ThinLTO function importing and other pre-optimization pipeline changes."? |
Can we do this clang invocation and diff later on so you don't need to redo the opt and llvm-lto steps below?